1361787 - Printing hang with box-shadow on windows 10

Status: VERIFIED FIXED

(Core :: Printing: Output, defect)

Description

[Bernard Leverd]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:53.0) Gecko/20100101 Firefox/53.0
Build ID: 20170413192749

Steps to reproduce:

Try to print page 1 of http://essentialseeing.com/phototrip/exclusive-namibia-namibrand-photo-workshop/
set fit to page if not set already
hit print


Actual results:

the progression bar goes to (apparently) 100% but stays there for ever
hitting cancel or closing the progression bar window : Firefox does not respond and has to  be killed.
In another test , I noticed in the windows spool list that the document size was more than 2 GB for  2 pages.
Printing from internet explorer is OK, from other applications is OK, but printing from Thunderbird 52 gives similar problem to Firefox

Comment 1

[Alice0775 White]

I can reproduce the issue on Windows10, but not on Ubuntu16.04.

And this seems not to be a recent regression.

Comment 2

[Alice0775 White]

Regression window:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=cbe9a2aea954&tochange=5ddf0a252b08

Regressed by:
69aca2db34d	Mason Chang — Bug 1162824 - Change box shadow cache to cache the colored blurred box shadow. r=mstange
28bbd1fb7ed1	Mason Chang — Bug 1155828 - Draw box-shadows using an approach inspired by border-image. r=mstange



And  *{box-shadow:none !important;} fixes the hang.

@:mchang,
Your bunch of patches causes the print hang. Please look this.

Comment 3

[Alice0775 White]

Attachment: possible minimum html

Firefox53 32 bit hangs when print with "Microsoft print to PDF" on Windows10 64bit.

Comment 4

[Mason Chang [Inactive] [:mchang]]

(In reply to Alice0775 White from comment #3)
> Created attachment 8864334 [details]
> possible minimum html
> 
> Firefox53 32 bit hangs when print with "Microsoft print to PDF" on Windows10
> 64bit.

This prints fine for me but the original page doesn't.

Comment 5

[Mason Chang [Inactive] [:mchang]]

(In reply to Mason Chang [:mchang] from comment #4)
> (In reply to Alice0775 White from comment #3)
> > Created attachment 8864334 [details]
> > possible minimum html
> > 
> > Firefox53 32 bit hangs when print with "Microsoft print to PDF" on Windows10
> > 64bit.
> 
> This prints fine for me but the original page doesn't.

This is odd. If I change the values of the box shadow to any other value, the page prints fine. Also, this test page crashes for me now.

Comment 6

[Mason Chang [Inactive] [:mchang]]

Can you try this build when it's done?

https://archive.mozilla.org/pub/firefox/try-builds/[e-mail]-379e2cb9259bcfa95588419ec62fd9255e7b9efc/

Comment 7

[Alice0775 White]

(In reply to Mason Chang [:mchang] from comment #6)
> Can you try this build when it's done?
> 
> https://archive.mozilla.org/pub/firefox/try-builds/[e-mail]-
> 379e2cb9259bcfa95588419ec62fd9255e7b9efc/

The try builds(try-win32-debug,try-win64-debug,try-win64) fixes the hang, I cannot reproduce the hang any more. And I can confirm that print result is fine (tested with URL page and possible minimum html)

(Of course, I can still reproduce the hang on latest Nightly55.0a1(2017-05-09) 32bit and 64bit.)

Comment 8

[Mason Chang [Inactive] [:mchang]]

Attachment: Don't use min box shadow when printing

Comment 9

[Mason Chang [Inactive] [:mchang]]

Try - https://treeherder.mozilla.org/#/jobs?repo=try&revision=02c116e80befa4185f8be7f3d73bcb70dba25be9

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/8569a2b1c1ee
Don't use the minimum box shadow when printing box shadows. r=lsalzman

Comment 11

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/8569a2b1c1ee

Comment 12

[Ryan VanderMeulen [:RyanVM]]

Sounds like this could use Beta & ESR52 approval requests.

Comment 13

[Mason Chang [Inactive] [:mchang]]

Comment on attachment 8866387 [details] [diff] [review]
Don't use min box shadow when printing

Approval Request Comment
[Feature/Bug causing the regression]: Bug 1155828 
[User impact if declined]: Printing can cause Firefox to crash
[Is this code covered by automated tests?]: Yes and no. Box shadows are but not printing.
[Has the fix been verified in Nightly?]: Yes
[Needs manual test from QE? If yes, steps to reproduce]: See comment 0.
[List of other uplifts needed for the feature/fix]: None
[Is the change risky?]: Maybe
[Why is the change risky/not risky?]: This isn't a big change, except we've had lots of long tail box shadow regressions before during Bug 1155828. We made reftests for those regressions when they came up, but I'm still wary.
[String changes made/needed]:

Comment 14

[Gerry Chang [:gchang]]

Comment on attachment 8866387 [details] [diff] [review]
Don't use min box shadow when printing

Fix a hang issue related to printing. Beta54+. Should be in 54 beta 8.

Comment 15

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-beta/rev/9ded11e5e796

Comment 16

[Paul Silaghi, QA [:pauly]]

Reproduced on 53.0.2.
Verified fixed Fx 54b8, 55.0a1 (2017-05-15) Win 10.

Comment 17

[Mason Chang [Inactive] [:mchang]]

If we don't have any new regressions from this by Friday, I guess we can ESR approval this.

Comment 18

[Mason Chang [Inactive] [:mchang]]

Comment on attachment 8866387 [details] [diff] [review]
Don't use min box shadow when printing

See https://bugzilla.mozilla.org/show_bug.cgi?id=1361787#c13

Comment 19

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Comment on attachment 8866387 [details] [diff] [review]
Don't use min box shadow when printing

I looked at the comments and the user impact, I am not convinced this is a widespread enough problem for ESR users. The uplift bar for ESR is generally severe new regressions, security and stability issues. As such, this fix doesn't seem so critical to take the risk of unknown regressions that this might induce.